Minimally invasive surgical techniques have been highly successful in improving patient care, reducing risk of infection, and decreasing recovery times. This project aims to further reduce invasiveness by developing a technique to insert thin, flexible needles into the human body and steer them from outside. This approach will potentially improve a wide range of procedures, from chemotherapy and radiotherapy to biopsy collection and tumor ablation, by enhancing physicians'abilities to accurately maneuver inside the human body without additional trauma to the patient. Building on emerging methods in robotics and highly encouraging results obtained under an R21 exploratory grant, we propose to design, prototype, and evaluate a working system that will steer flexible needles through deformable tissue and around internal obstacles to precisely reach specified 3D targets. This research program will significantly advance our understanding and practice of needle therapies through integrated needle design and modeling, preoperative visualization and needle motion planning, and image-guided intraoperative needle control. The scientific and engineering advances will culminate in a set of pre-clinical trials with imaging (fluoroscopy, ultrasound, and MRI) using phantom and natural ex vivo and in vivo models. The designs, analyses, and experiments of this study will determine the merits and weaknesses of flexible needle steering, with the goals of improving current clinical applications and leading to new ultra-minimally invasive surgical procedures. The results of this project could significantly improve public health by lowering patient recovery times, infection rates, and treatment costs. By increasing the dexterity and accuracy of minimally invasive procedures, anticipated results will not only improve outcomes of existing procedures, but also enable percutaneous procedures for many conditions that currently require open surgery.